Академический Документы
Профессиональный Документы
Культура Документы
e-ISSN: 2278-5728, p-ISSN: 2319-765X. Volume 11, Issue 4 Ver. V (Jul - Aug. 2015), PP 47-61
www.iosrjournals.org
Associate professor, Department of Mathematics, Patuakhali Science and Technology University, Dumki,
Patuakhali, Bangladesh
2
Associate professor, Department of Computer Science and Information Technology, Patuakhali Science and
Technology University, Dumki, Patuakhali, Bangladesh
3
Associate professor, Department of Statistics, Patuakhali Science and Technology University, Dumki,
Patuakhali, Bangladesh
Abstract: In this paper, we present the solution of one dimensional advection diffusion equation for initial
condition in infinite space analytically by transform to heat equation via coordinate transformation. A
comparison among explicit upwind difference scheme, explicit centered difference scheme and explicit
downwind difference scheme is projected herein with a variety of numerical results and relative errors. Our
goal is to investigate the efficient numerical schemes for advection diffusion equation.
Keywords: Advection Diffusion Equation, Explicit Scheme, Relative Errors.
I.
Introduction
The mathematical model describing the transport and diffusion processes is the one dimensional
2
advection-diffusion equation (ADE): C u C D C with constant coefficients. Mathematical modeling of
2
heat transport, pollutants, and suspended matter in water and soil involves the numerical solution of an
Advection diffusion equation. Many researchers are involved for solving the model equation (ADE) by using
the finite difference scheme. Agusto and Bamigbola (2007) studied on the Numerical treatment of the
mathematical model for water pollution. This study was examined by various mathematical models involving
water pollutant. The authors used the implicit centered difference scheme in space and a forward difference
method in time for the evaluation of the generalized transport equation. Kumar et al (2009) presented an
analytical solution of one dimensional advection diffusion equation with variable coefficients in a finite domain
using Laplace transformation technique. The authors introduced new independent space and time variables in
this process. In this study the analytical solution was compared with the numerical solution in case the
dispersion is proportional to the same linearly interpolated velocity. Kumar et al (2011) made an Analytical
solution of the advection diffusion equation with temporally dependent Coefficients. Park et al (2008)
performed an analytical solution of the advection diffusion equation for a ground level finite area source using
superposition method. Van Genuchten et al studied an Analytical solution of the advection diffusion transport
equation using a change of variable and integral transform technique. Thongmoon and Mckibbin (2006)
compared some Numerical Methods for the Advection-Diffusion Equation. The authors reported that the finite
difference methods (FTCS, Crank Nicolson) give better point-wise solutions than the spline methods. Yuste et al
(2005) described an explicit finite difference method and a new Von Numann-type stability analysis for
fractional diffusion equations. Changiun Zhu et al (2010) conducted a study on a Numerical Simulation of
Hybrid Finite Analytic Method for Ground Water Pollution.
Therefore, in section II, we present the derivation of advection diffusion equation on the principle of
conservation law of mass using Ficks law based on ([1], [2], [3], [7]). In section III, we solve one dimensional
advection diffusion equation for initial condition in infinite space analytically by transform to heat equation via
coordinate transformation based on ([1], [5], [11]). Based on the study of the general finite difference method
for the second order linear partial differential equation ([7], [14], [16], [17]) we develop a explicit finite
difference scheme for ADE as an IBVP with two sided boundary conditions in section IV. In this section, we
also establish the stability condition of the explicit centered difference scheme for advection diffusion equation.
In section V, we present an algorithm for the numerical solution and we implement the numerical scheme in
order to perform the numerical features of error estimation. Our goal is to investigate the efficient numerical
schemes for advection diffusion equation. Finally the conclusions of the paper are given in last section.
DOI: 10.9790/5728-11454761
www.iosrjournals.org
47 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
II.
Many partial differential equation models with a physical motivation develop from a conservation law.
A conservation law is just a mathematical formulation of the basics fact that the rate at which a quantity changes
in a given domain must equal the rate at which the quantity flows across the boundary plus the rate at which the
quantity is created or destroyed within the domain. Let C C ( x, t ) denote the density of a given quantity
(energy, mass, Automobiles, etc), density is usually measured in amount per unit volume or some time amount
per unit length. Further, we let q q( x, t ) denote the flux of the quantity crossing the section at x and at time
and its unit are given in amount per unit area, per unit time. Thus,
quantity that is crossing the section at x at time t . Where A is the cross sectional area of the tube. Finally
f f ( x, t ) denote the given rate at which the quantity is created or destroyed within the section at x at
time t . The function f is called the source term if it is positive, and a sink if it is negative, it is measured in
amount per unit volume per unit time. We can derive the law by assuming a fixed time but arbitrary, section
[ x1 , x2 ] of the tube and requiring that the rate of change of the total amount of the quantity in the section must
equal the rate at which it flows in at x x1 , minus the rate at which it flows out at x x2 , plus the rate at
which it is created within [ x1 , x2 ] .
In mathematically, we can define the conservation law.
x
2
d 2
C
(
x
,
t
)
Adx
Aq
(
x
,
t
)
Aq
(
x
,
t
)
1
2
f ( x, t )Adx
dt x1
x1
(1)
This is one integral form of conservation law. Here A is constant, it may be cancelled from the formula.
However, if the function C and q are sufficiently smooth, then it may reformulated as a PDE model.
x
2
d 2
C
(
x
,
t
)
dx
q
(
x
,
t
)
q
(
x
,
t
)
1
2
f ( x, t )dx
dt x1
x1
x2
x2
x2
Ct ( x, t )dx q x ( x, t ) f ( x, t )dx
x1
x1
x2
x2
C ( x, t )dx q
t
x1
x1
x2
( x, t ) f ( x, t )dx 0
x1
x1
Ct ( x, t ) q x ( x, t ) f ( x, t ) 0
Ct ( x, t ) q x ( x, t ) f ( x, t )
(2)
qnet q J
= uc D
c
x
We substitute this quantity at the conservation law (2) in the absence of source. And the conservation law
becomes
DOI: 10.9790/5728-11454761
www.iosrjournals.org
48 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
C
)x 0
x
C
C
2C
u
D 2 0
t
x
x
C
C
2C
u
D 2
t
x
x
Ct ( x, t ) (uC D
J xdirection uc J
uc D
Where,
c
x
uc
in
out
x -direction, we have
c
c
) yz (uc D ) yz
x 1
x 2
We use linear Taylor series expansion to combine the two flux terms, giving
m x (uc D
uc 1 uc 2 uc 1 (uc 1
(uc)
x)
x 1
(uc)
x
x
DOI: 10.9790/5728-11454761
www.iosrjournals.org
49 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
And
c
c
D
x 1
x
D
2
2c
x
x 2
c
c
c
(D
( D ) x)
x 1
x 1 x
x 1
x -direction
m x
(cu )
2c
xyz D 2 xyz
x
x
m y
(cv )
c
yxz D 2 yxz
y
y
m z
(cw)
2c
zxy D 2 zxy
z
z
M cxyz
We obtain
c
.(cu ) D 2 c
t
or in Einsteinian notation
c cu i
2c
D 2
t xi
xi
This is the desired advection diffusion equation (ADE).
In the one-dimensional case, u (u,0,0) and there are no concentration gradients in the y -direction or
z -direction, leaving us with
Since
c uc
2c
D 2
t x
x
is constant, then
c
c
2c
u D 2
t
x
x
(4)
www.iosrjournals.org
50 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
This coordinate system, essence, translates any given fixed location x , to the distance between fixed
location and observers location, which is described by ut . In addition, we will let t and then C ( x, t )
becomes C ( , ) . Using the chain rule, we have
c c c c
x x x
2c 2c
and
x 2 2
Similarly, the time derivative term can be expressed as
c c c
c c
t t t
Substituting of the above expression to the advection diffusion equation leads to the following expression
c c
c
2c
u
u
D 2
III.
With
c
2c
D 2
(7)
c(0, ) c0 ( )
(8)
Driving the solution of (7)-(8) is accomplished in two steps. First we will solve the problem for a
special step function c0 ( ) , and then we will construct the solution to (7)-(8) using the special solution
So firstly we consider the problem
c
2c
D 2
(9)
c(0, ) c0 for 0
and c(0, ) 0 for 0
With
(10)
Where we have taken the initial condition to be a step function with jump c0
3.1 The fundamental Solution of the Heat Equation
We persuade our approach to the solution (9)-(10) with a simple idea from the subject of dimensional
analysis. Dimensional analysis deals with the study of units (seconds, meters, kilogram, and so forth) and
dimensions (time, length, mass and so forth) of the quantities in a problem and how relate to each other.
Equations must be dimensionally consistent (one cannot add apples or oranges), and important conclusions can
be drawn from this fact. The cornerstone result in dimensional analysis is called the pi theorem. The pi theorem
guarantees that whenever there is a physical law dimensionless quantities q1 , q2 , q3 ,............, qm then there is
an equivalence physical law relating the independent dimensionless quantities that can be formed
from q1 , q2 , q3 ,............, qm . By a dimensional quantity we mean one in which all the dimensions (time, length,
mass and so forth) cancel out. As an example take the law
DOI: 10.9790/5728-11454761
www.iosrjournals.org
51 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
1
h gt 2 ut
2
That gives the height h of an object at time t when something is thrown upwind with initial
velocity u ; the constant g is the acceleration due to gravity. Here the dimensioned quantities are h, t , u and g ,
having dimensions length, time, length per time, and length per time squared. This law can be rearranged and
written equivalently as
h
gt
1
ut
2u
h and
gt
1
2
ut
u
For example, h is length and ut , also a length. So 1 has no dimensions and similarly 2 is
dimensionless.
We use similar reasoning to guess the form of the IVP (9)-(10). First we list all the variables and
constants in the problem , , c, c0 , D . These have dimensions length, time, degrees, and length squared per
time respectively. We notice that
c
is a dimensionless quantity (degrees divided by degrees), the only other
c0
4 D
We accept therefore that the solution can be written as some combination of these dimensionless
variables, or c f ( ) for some function f to be determined.
c0
4 D
So let us substitute c f (z ) , z
into the PDE (9). We have taken c0 1 for simplicity. The chain
4 D
rule allows us to compute the partial derivatives as
1
c f ( z ) z
f ( z )
2 4 D 3
1
f ( z )
4 D
c f ( z ) z
c
1
1
1
f ( z )
f ( z )
4
D
4 D
4 D
1
1 1
f ( z ) z f ( z )
4
2
1
1 1
f ( z ) z f ( z )
4
2
f ( z) 2 zf ( z)
f ( z) 2 zf ( z) 0 , for f (z )
This is an ODE
z2
Here integrating factor is e . The equation is easily solved by multiplying through by integrating
factor e
We get
z2
and integrating.
2
f ( z ) k1 e r dr k 2 , where k 2 is another integrating constant.
www.iosrjournals.org
52 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
c( , ) k1
4 D
r 2
dr k 2 ,
To determine the constant k1 and k 2 we apply the initial condition. For a fixed
to get
0 c( ,0) k1 e r dr k 2 ,
2
1 c( ,0) k1 e r dr k 2
2
Recalling that
r 2
dr
1 and k 1
2
c0 1 is c( , ) 1
4 D
r 2
dr
1
2
This solution can be presented by the special function called the error function as
c( , ) (1 erf (
))
2
4 D
erf ( z )
r 2
dr
Firstly if a function c satisfies the heat equation, then the function c satisfies the heat equation.
0 (c Dc ) (c ) D(c ) (c ) D(c )
{Since
Hence
(c )
1
1
1
[ f ( z ) zf ( z )] and (c )
3
2
2 4 D
(c ) (c )
1
4 D 3
[ f ( z ) zf ( z )]
(11)
DOI: 10.9790/5728-11454761
www.iosrjournals.org
53 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
IV.
In this section, the study investigates a finite difference scheme for the water pollution model as
a parabolic second order partial differential equation. This chapter contains the analysis of the condition
of stability of the explicit finite difference scheme.
4.1 Explicit Upwind difference schemes for Advection Diffusion Equation
We consider our second order water pollution model
c
c
2c
u
D 2
t
x
x
(12)
O(t )
t
t
c
The discreetization of
is obtained by first order backward difference in space
x
The discreetization of
c Cin Cin1
O(x)
x
x
The discretization of
2c
is obtain from second order centered difference in space.
x 2
2 c Cin1 2Cin Cin1
O(x 2 )
x 2
x 2
cin1 cin
c n cin1
c n 2cin cin1
u i
D i1
t
x
x 2
c n cin1
c n 2cin cin1
cin1 cin ut i
Dt i1
x
x 2
=> cin1 (
Dt ut n
ut
Dt
Dt
)ci1 (1
2 2 )cin 2 cin1
2
x
x
x
x
x
(13)
where u t , D t
x
x 2
Which is the explicit upwind difference scheme and it is also known as FTBSCS techniques. The
t
t
stability condition is controlled by u
,D
x
x 2
4.2 Explicit centered difference scheme for Advection Diffusion Equation
We consider our second order water pollution model
c
c
2c
u
D 2
t
x
x
(14)
c Cin1 Cin
O(t )
t
t
Centered difference discretization in spatial derivative:
DOI: 10.9790/5728-11454761
www.iosrjournals.org
54 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
c Cin1 Cin1
O(x 2 )
x
2x
Discretization of
2c
is obtain from second order centered difference in space.
x 2
2c Cin1 2Cin Cin1
O(x 2 )
2
2
x
x
cin1 cin
cn cn
c n 2cin cin1
u i 1 i 1 D i 1
t
2x
x 2
n 1
=> ci
Dt ut n
Dt
Dt ut n
)ci 1 (1 2 2 )cin ( 2
)ci 1
2
2x
2x
x
x
x
(15)
c
c
2c
u
D 2
t
x
x
Let the solution
c( xi , t n ) be denoted by C in
The discreetization of
The discreetization of
(16)
n
c
is obtained by first order forward difference in time
t
c Cin1 Cin
O(t )
t
t
c
is obtained by first order forward difference in space
x
c Cin1 Cin
O(x)
x
x
The discretization of
2c
is obtain from second order centered difference in space.
x 2
2c Cin1 2Cin Cin1
O(x 2 )
2
2
x
x
cin1 cin
cn cn
c n 2cin cin1
u i1 i D i1
t
x
x 2
=>
cin1
Dt n
ut
Dt
Dt ut n
c (1
2 2 )cin ( 2
)ci1
2 i 1
x
x
x
x
x
www.iosrjournals.org
(17)
55 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
Which is the explicit downwind difference scheme and it is also known as FTFSCS techniques. The stability
t
t
, D
x
x 2
condition is controlled by u
Now we can write the general form of an explicit finite difference scheme of advection diffusion equation as
Lemma 4.3-1: Stability of the explicit centered difference scheme (15) of Advection Diffusion equation is
given by the conditions
t
0u
1 , 0 D t2 1
x
2
x
Proof: The explicit centered difference scheme for (14) is given
cin1 (
n 1
=> ci
where
Dt ut n
Dt
Dt ut n
)ci 1 (1 2 2 )cin ( 2
)ci 1
2
2x
2x
x
x
x
t
,
x
(18)
t
x 2
0 1
2
0 1 2 1
0 1
2
(i)
(ii)
(iii)
The new solution is a convex combination of the two previous solutions. That is the solution at new
time-step (n 1) at a spatial node i is an average of the solutions at the previous time-step at the spatialnodes i 1 , i and i 1 . This means that the extreme value of the new solution is the average of the extreme
values of the previous two solutions at the three consecutive nodes.
In our model the characteristics speed u is assumed to positive.
Then we have
From equation (ii), 0 1 2 1
ut
0
x
1 2 1 1
1
0
2
www.iosrjournals.org
56 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
0 1
2
1
0 1
2 2
It is clear that 1
We can conclude that the explicit centered difference scheme (18) is stable for
Where 0 u
t
t
1
1 and 0 D 2
x
2
x
V.
To find out the numerical solution of the model, we have to accumulate some variables which are
offered in the following algorithm.
Input:
nx
and
nt
dx
b0
, the spatial grid size
nx
gm u *
Step1. Calculation for numerical solution of ADE by explicit centered difference scheme
For n 1 to nt
For i 2 to nx
end
end
Step2; Output c( x, t )
Step3: Figure Presentation
Step4: Stop
DOI: 10.9790/5728-11454761
www.iosrjournals.org
57 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
e1
Ce C N
Ce
for all time where C e is the exact solution and C N is the Numerical solution computed by the explicit finite
difference scheme.
t 60 in
DOI: 10.9790/5728-11454761
www.iosrjournals.org
58 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
Figure 5.4: Comparison of Relative errors between explicit upwind difference scheme and explicit
centered difference scheme
DOI: 10.9790/5728-11454761
www.iosrjournals.org
59 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
Figure 5.5: Comparison of Relative errors between explicit centered difference scheme and explicit
downwind difference scheme
5.1.5 Comparison of explicit upwind difference scheme, explicit centered difference scheme and explicit
downwind difference scheme
Figure 5.6: Comparison of Relative errors among explicit upwind difference scheme, explicit centered
difference scheme and explicit downwind difference scheme
Figure (Figure 5.1-5.6) shows the relative error of three different schemes for advection diffusion
equation. Figure 5.1 shows the relative error for EUDS, which remains below 0.0212. Figure 5.2 shows the
relative error for ECDS which remains below 0.0110. Figure 5.3 shows the relative error for EDDS scheme
which remains below 0.0401. ECDS provides accurate results than the EUDS and EDDS scheme with respect to
discretization parameter x .
VI.
Conclusion
The study has presented the numerical and analytical solution of Advection Diffusion equation. We
have studied three explicit differences schemes for the numerical solutions of Advection Diffusion Equation. It
has also compared among the three different schemes; EUDS, ECDS and EDDS scheme for estimating the
relative error in advection diffusion equation. The explicit centered difference scheme is more efficient
numerical scheme for ADE. The explicit centered difference scheme can be extended for two dimensional
advection diffusion equations as a water pollution model which demands the further study.
DOI: 10.9790/5728-11454761
www.iosrjournals.org
60 | Page
Error Estimation Of An Explicit Finite Difference Scheme For A Water Pollution Model
Bibliography
[1].
[2].
[3].
[4].
[5].
[6].
[7].
[8].
[9].
[10].
[11].
[12].
[13].
[14].
[15].
[16].
[17].
[18].
[19].
DOI: 10.9790/5728-11454761
www.iosrjournals.org
61 | Page